The Uninsurance Rate Held Steady during the Pandemic as Public Coverage Increased
←
→
Page content transcription
If your browser does not render page correctly, please read the page content below
HEALTH POLICY CENTER
Support for this research was
provided by the Robert Wood
Johnson Foundation. The views
expressed here do not necessarily
reflect the views of the
Foundation.
The Uninsurance Rate Held Steady during
the Pandemic as Public Coverage Increased
Trends in Health Insurance Coverage between March 2019 and April 2021
Michael Karpman and Stephen Zuckerman
August 2021
Rapid job losses in the early months of the COVID-19 pandemic raised fears that
millions of people would lose their health insurance coverage and become uninsured
(Banthin et al. 2020; Garfield et al. 2020; Garrett and Gangopadhyaya 2020). In previous
recessions, laid-off workers who lost employer-sponsored insurance (ESI) faced limited
coverage options through Medicaid and the private nongroup insurance market and the
number of people uninsured increased (Holahan and Chen 2011). The Affordable Care
Act (ACA) significantly expanded access to those options in 2014, driving the
uninsurance rate to record lows (ASPE 2021; Obama 2016). And as the pandemic posed
the first test of the post-ACA health insurance safety net during an economic downturn,
Congress further supported access to coverage by not allowing disenrollment from
Medicaid through the March 2020 Families First Coronavirus Response Act (Brooks and
Schneider 2020).1
In this brief, we examine changes in health insurance coverage among nonelderly adults ages 18 to
64 during the pandemic using data from the Urban Institute’s Health Reform Monitoring Survey
(HRMS). Since it was launched in 2013, the HRMS has provided timely information on coverage before
data from federal surveys become available (Long et al. 2014). Our analysis focuses on changes in
coverage across three rounds of the survey: March 2019; March/April 2020, just after the pandemic
caused a steep decline in employment; and April 2021, more than one year after the secretary of
health and human services declared a national public health emergency on January 31, 2020. We
estimate regression-adjusted changes for the national nonelderly adult population overall, by state
Medicaid expansion status,2 and by annual family income as a percentage of the federal poverty level(FPL). We focus on adults with low incomes targeted by the ACA Medicaid expansion (with incomes at
or below 138 percent of FPL) and adults with moderate incomes eligible for ACA Marketplace
premium tax credits (with incomes between 139 and 399 percent of FPL).3 We find the following:
◼ Between March 2019 and April 2021, the share of nonelderly adults reporting ESI declined
from 65.0 to 62.3 percent, a decrease of approximately 5.5 million adults. The share reporting
public coverage increased from 13.6 to 17.5 percent, an increase of approximately 7.9 million
adults. The national uninsurance rate held steady at approximately 11 percent.
◼ The share of adults reporting public coverage increased between 2019 and 2021 in both
states that had and had not expanded Medicaid under the ACA (hereafter called expansion
and nonexpansion states). Such coverage increased from 14.9 to 19.2 percent in expansion
states and from 10.7 to 14.3 percent in nonexpansion states.
◼ In Medicaid expansion states, the uninsurance rate was near 8 percent across all three study
years. In nonexpansion states, the uninsurance rate was higher in 2021 (18.2 percent) than in
2020 (16.5 percent) and 2019 (17.2 percent), though the difference between 2019 and 2021
was not statistically significant. Adults in nonexpansion states were more than twice as likely
as adults in expansion states to be uninsured in 2021 (18.2 percent versus 7.7 percent).
◼ Declines in ESI and increases in public coverage between 2019 and 2021 were concentrated
among adults with low and moderate incomes. Uninsurance rates among the national
nonelderly adult population did not change significantly for any income group examined.
◼ The share of adults with low incomes reporting public coverage increased in both expansion
states (from 54.6 to 62.9 percent) and nonexpansion states (from 30.4 to 37.3 percent)
between 2019 and 2021. More than one in three adults with low incomes in nonexpansion
states (37.7 percent) were uninsured in 2021, compared with about one in seven of such
adults in expansion states (14.5 percent).
Between 2019 and 2021, the rise in public coverage helped offset a decline in ESI, and unlike in
previous recessions, the uninsurance rate did not change. Medicaid and, to a lesser extent, private
nongroup insurance sold through the Marketplaces have provided many adults with coverage options
following unprecedented job and income losses. However, more than 1 in 10 adults were uninsured in
April 2021, including nearly 1 in 5 adults in nonexpansion states.
Maintaining the current uninsurance rate will require protecting coverage for current and
prospective Medicaid enrollees as the economy improves and the disenrollment freeze is lifted (which
is unlikely to occur before early 2022). Adults eligible for Medicaid may be at risk of having their
applications or renewals erroneously rejected if states resume normal operations for reviewing
eligibility too rapidly (Rosenbaum, Handley, and Morris 2021). Other adults will no longer be eligible
for Medicaid when their incomes recover and will need to seek private coverage to remain insured.
For those without access to affordable ESI, outreach efforts can raise their awareness of the enhanced
premium tax credits for Marketplace plans made available under the March 2021 American Rescue
Plan Act (Haley and Wengle 2021). States will also need to assess eligibility for subsidized Marketplace
2 UNINSURANCE HELD STEADY , PUBLIC COVERAGE INCREASED DURING THE PANDEMICcoverage for people losing Medicaid eligibility after the public health emergency ends (Musumeci and
Dolan 2021). Permanently extending the American Rescue Plan Act’s enhanced tax credits could
further reduce the number of uninsured people over the long term, and adults with moderate incomes
would experience the largest decline in uninsurance (Banthin et al. 2021). Policymakers can also build
on coverage gains under the ACA by addressing the persistently high uninsurance rates among adults
with low incomes, particularly in nonexpansion states.
Results
Between March 2019 and April 2021, the share of nonelderly adults reporting ESI declined and the share
reporting public coverage increased; the national uninsurance rate held steady.
Approximately 65 percent of nonelderly adults reported having ESI coverage in March 2019 and
March/April 2020 (figure 1).4 This share had declined to 62.3 percent by April 2021, when many
adults remained out of work just over one year after the pandemic recession began.5 The 2.7
percentage-point decline in ESI between 2019 and 2021 represents a decrease of approximately 5.5
million adults (95 percent confidence interval: 2.5 million, 8.5 million).6 During this period, the share of
adults reporting public coverage—including Medicare, Medicaid, the Children’s Health Insurance
Program (CHIP), and other state or government plans based on income or disability7—increased from
13.6 percent in 2019 to 17.5 percent in 2021, representing an increase of approximately 7.9 million
adults (95 percent confidence interval: 5.4 million, 10.4 million).8
We did not observe a statistically significant change in private nongroup coverage, which
approximately 8 percent of adults reported in each year and includes plans purchased through and
outside the ACA Marketplaces.9 But the share of adults with unspecified coverage (i.e., reporting the
name of a comprehensive health plan but not the type of coverage) declined by 1.1 percentage points
between 2019 and 2021.10 The share of adults with unspecified coverage was also slightly higher in
2019 than in March 2018, suggesting an anomalous result in 2019 (data not shown). Despite the
significant loss of ESI, the uninsurance rate held steady nationally at approximately 11 percent in each
study year.
Net changes in ESI, public coverage, and private nongroup coverage do not fully capture the
transitions across coverage types that may have occurred during the pandemic. Income losses made
some adults eligible for Medicaid and others eligible for subsidized Marketplace coverage, regardless
of whether they were previously covered by ESI. The lack of net change in nongroup coverage could
indicate that new Marketplace enrollment among people who became eligible for premium tax credits
was not large enough to offset transitions from Marketplace or non-Marketplace nongroup coverage
to Medicaid. In addition, the sample size of the HRMS may not be large enough to detect statistical
significance for the relatively small changes in Marketplace enrollment found in administrative data.
UNINSURANCE HELD STEADY, PUBLIC COVERAGE INCREASED DURING THE PANDEMIC 3FIGURE 1
Health Insurance Coverage among Adults Ages 18 to 64, March 2019 to April 2021
Percent
March 2019 March/April 2020 April 2021
65.0 65.1
62.3***^^^
17.5***^^^
13.6 15.0***
11.1 11.0 11.2
8.0 7.5 7.7
ESI Public coverage Private nongroup coverage Uninsured
URBAN INSTITUTE
Source: Health Reform Monitoring Survey, March 2019 through April 2021.
Notes: ESI is employer-sponsored insurance. Estimates are regression adjusted. Estimates are not shown for the share of adults
with an unspecified coverage type (2.3 percent in 2019, 1.4 percent in 2020, and 1.3 percent in 2021).
*/**/*** Estimate differs significantly from that for March 2019 at the 0.10/0.05/0.01 level, using two-tailed tests.
^/^^/^^^ Estimate differs significantly from that for March/April 2020 at the 0.10/0.05/0.01 level, using two-tailed tests.
The share of adults reporting public coverage increased in both Medicaid expansion and nonexpansion
states.
As shown in figure 2, ESI coverage declined between 2019 and 2021 in expansion states (from 67.0 to
64.6 percent) and nonexpansion states (from 61.3 to 57.9 percent). But public coverage increased
during this period in both groups of states, from 14.9 to 19.2 percent in expansion states and from
10.7 to 14.3 percent in nonexpansion states. These patterns are consistent with Centers for Medicare
& Medicaid Services data showing rapid Medicaid enrollment growth in both expansion and
nonexpansion states during the pandemic (Corallo and Rudowitz 2021; Khorrami and Sommers
2021).11
The higher rates of public coverage in expansion states than in nonexpansion states in both 2019
and 2021 largely reflect the former’s more generous eligibility for Medicaid; nearly all adults living in
expansion states with incomes below 138 percent of FPL are eligible.12 In nonexpansion states,
nondisabled, nonpregnant parents typically must have very low incomes to qualify for Medicaid (e.g.,
17 percent and 18 percent of FPL in Texas and Alabama) and nonparents are ineligible. 13 The increase
in reported public coverage in nonexpansion states over the study period was concentrated among the
groups most likely to be eligible for Medicaid or CHIP.14
The uninsurance rate in Medicaid expansion states was approximately 8 percent between 2019
and 2021. In nonexpansion states, the uninsurance rate was higher in 2021 (18.2 percent) than in
4 UNINSURANCE HELD STEADY , PUBLIC COVERAGE INCREASED DURING THE PANDEMIC2020 (16.5 percent) and 2019 (17.2 percent), though the difference between 2019 and 2021 was not
statistically significant. As in prior years, adults in nonexpansion states were more than twice as likely
as adults in expansion states to be uninsured in 2021 (18.2 versus 7.7 percent). However, differences
in uninsurance are not entirely attributable to differences in Medicaid eligibility, because other factors
(e.g., access to ESI, funding for outreach and enrollment assistance) likely affect coverage status.
FIGURE 2
Health Insurance Coverage among Adults Ages 18 to 64, by State Medicaid Expansion Status,
March 2019 to April 2021
Expansion states (%)
March 2019 March/April 2020 April 2021
67.0 66.9 64.6***^^^
19.2***^^^
14.9 16.3**
7.7 7.0 7.3 8.0 8.2 7.7
ESI Public coverage Private nongroup coverage Uninsured
Nonexpansion states (%)
61.3 61.4
57.9**^^^
17.2 16.5 18.2^^
10.7 12.4** 14.3***^
8.5 8.6 8.5
ESI Public coverage Private nongroup coverage Uninsured
URBAN INSTITUTE
Source: Health Reform Monitoring Survey, March 2019 through April 2021.
Notes: ESI is employer-sponsored insurance. Medicaid expansion states implemented expansions by April 2021. Estimates are
regression adjusted. Estimates are not shown for the share of adults with an unspecified coverage type; these shares were 2.4
percent, 1.6 percent, and 1.3 percent in expansion states in 2019, 2020, and 2021 and 2.3 percent, 1.1 percent, and 1.2 percent
in nonexpansion states in 2019, 2020, and 2021.
*/**/*** Estimate differs significantly from that for March 2019 at the 0.10/0.05/0.01 level, using two-tailed tests.
^/^^/^^^ Estimate differs significantly from that for March/April 2020 at the 0.10/0.05/0.01 level, using two-tailed tests.
UNINSURANCE HELD STEADY, PUBLIC COVERAGE INCREASED DURING THE PANDEMIC 5Declines in ESI and increases in public coverage between 2019 and 2021 were concentrated among adults
with low and moderate incomes.
Adults with low and moderate incomes were hardest hit by the recession (Karpman, Zuckerman, and
Kenney 2020)15 and reported the largest declines in ESI over the study period. Among adults with
past-year incomes at or below 138 percent of FPL, the share with ESI fell from 21.4 to 16.0 percent
during this period (table 1). Among adults with incomes between 139 and 399 percent of FPL, the
share with ESI fell from 64.5 to 60.0 percent. We did not find a statistically significant change in ESI
among adults with incomes at or above 400 percent of FPL.
Increased public coverage among adults with low incomes, from 45.0 to 52.6 percent, and those
with moderate incomes, from 9.7 to 14.3 percent, helped offset declines in ESI among these groups.
Most adults must have incomes below 138 percent of FPL to qualify for Medicaid in expansion states,
and eligibility in nonexpansion states is limited to parents with even lower incomes and generally
nonexistent for nonparent adults. However, eligibility is based on current monthly income, meaning an
adult whose annual family income in the past year was above the eligibility threshold may qualify if
they experience a loss of income that places them below the threshold.
The uninsurance rate did not change significantly in any of the income groups examined. Nearly
one in four adults with low incomes (23.7 percent) and about one in eight with moderate incomes
(12.8 percent) were uninsured in April 2021.
TABLE 1
Health Insurance Coverage among Adults Ages 18 to 64, by Family Income,
March 2019 to April 2021
Percent
Family income March 2019 March/April 2020 April 2021
At or below 138% of FPL
ESI 21.4 21.5 16.0***^^^
Public coverage 45.0 48.5** 52.6***^^
Private nongroup coverage 6.8 5.5 5.6
Uninsured 24.3 22.4 23.7
139–399% of FPL
ESI 64.5 64.0 60.0***^^^
Public coverage 9.7 10.8 14.3***^^^
Private nongroup coverage 11.3 10.8 11.8
Uninsured 11.8 12.9 12.8
At or above 400% of FPL
ESI 86.9 88.1 87.8
Public coverage 1.5 1.6 2.1**
Private nongroup coverage 6.1 5.8 5.3
Uninsured 3.6 3.5 3.7
Source: Health Reform Monitoring Survey, March 2019 through April 2021.
Notes: FPL is federal poverty level. ESI is employer-sponsored insurance. Estimates are regression adjusted. Estimates are not
shown for the share of adults with an unspecified coverage type, which is between 1 and 3 percent across income groups and
years.
*/**/*** Estimate differs significantly from that for March 2019 at the 0.10/0.05/0.01 level, using two-tailed tests.
^/^^/^^^ Estimate differs significantly from that for March/April 2020 at the 0.10/0.05/0.01 level, using two-tailed tests.
6 UNINSURANCE HELD STEADY , PUBLIC COVERAGE INCREASED DURING THE PANDEMICThe share of adults with low incomes reporting public coverage increased in both Medicaid expansion and
nonexpansion states between 2019 and 2021. More than one in three adults with low incomes in
nonexpansion states were uninsured in 2021, compared with about one in seven of such adults in expansion
states.
Among adults with incomes at or below 138 percent of FPL, the share reporting public coverage
increased from 54.6 to 62.9 percent in Medicaid expansion states and from 30.4 to 37.3 percent in
nonexpansion states between 2019 and 2021 (table 2). The uninsurance rate for adults with low
incomes was statistically unchanged in both groups of states, but wide disparities by Medicaid
expansion status persisted. In 2021, more than one in three adults with low incomes (37.7 percent) in
nonexpansion states were uninsured, compared with about one in seven (14.5 percent) of such adults
in expansion states. Adults with moderate incomes in nonexpansion states were nearly twice as likely
as those in expansion states to be uninsured (17.8 versus 10.1 percent).
TABLE 2
Health Insurance Coverage among Adults Ages 18 to 64, by State Medicaid Expansion Status and
Family Income, March 2019 to April 2021
Percent
Expansion States Nonexpansion States
March/ March/
March April March April
Family income 2019 2020 April 2021 2019 2020 April 2021
At or below 138% of FPL
ESI 20.7 21.2 15.7***^^^ 22.7 22.2 16.2*^^^
Public coverage 54.6 57.3 62.9***^^^ 30.4 34.3 37.3**
Private nongroup coverage 5.3 3.7** 4.4 8.8 8.3 7.6
Uninsured 16.5 15.0 14.5 36.3 34.3 37.7
139–399% of FPL
ESI 65.4 64.8 61.2**^^^ 63.1 62.6 57.7***^^
Public coverage 11.4 12.2 16.6***^^^ 6.2 8.0* 10.3***^
Private nongroup coverage 11.5 10.7 11.3 10.9 10.8 12.9
Uninsured 9.3 10.8 10.1 16.8 16.8 17.8
At or above 400% of FPL
ESI 87.9 88.3 88.4 84.6 87.3** 86.4
Public coverage 1.3 1.5 2.1*** 2.2 1.8 2.5
Private nongroup coverage 5.9 5.7 5.5 6.5 6.1 4.9*
Uninsured 2.9 3.2 3.0 4.9 4.2 5.3
Source: Health Reform Monitoring Survey, March 2019 through April 2021.
Notes: FPL is federal poverty level. ESI is employer-sponsored insurance. Medicaid expansion states implemented expansions by
April 2021. Estimates are regression adjusted. Estimates are not shown for the share of adults with an unspecified coverage
type, which is between 0 and 3 percent across income levels, state groups, and years.
*/**/*** Estimate differs significantly from that for March 2019 at the 0.10/0.05/0.01 level, using two-tailed tests.
^/^^/^^^ Estimate differs significantly from that for March/April 2020 at the 0.10/0.05/0.01 level, using two-tailed tests.
UNINSURANCE HELD STEADY, PUBLIC COVERAGE INCREASED DURING THE PANDEMIC 7Discussion
Despite losses of jobs, income, and ESI during the pandemic, the uninsurance rate did not change
between March 2019 and April 2021. Increased public coverage helped counter ESI losses, protecting
many adults from becoming uninsured both in Medicaid expansion and nonexpansion states. But in
April 2021, the uninsurance rate in nonexpansion states was higher than it had been in March/April
2020 and was more than double the uninsurance rate in expansion states.
The growth in public coverage reflects several factors, including expanded Medicaid eligibility
under the ACA that has strengthened the safety net in 37 states and the District of Columbia, the
freeze on Medicaid disenrollment under the Families First Coronavirus Response Act, and the historic
pattern of rising Medicaid enrollment during recessions (Corallo and Rudowitz 2021).16 Assessing how
each factor has affected coverage during the pandemic is beyond the scope of this brief. However, the
study findings highlight several challenges and opportunities for protecting and expanding coverage in
the near term.
Though the public health emergency and Medicaid disenrollment freeze will likely be extended at
least until early 2022,17 states will need to process a backlog of coverage renewals and
redeterminations when the freeze is lifted (Musumeci and Dolan 2021). Resuming normal operations
too quickly could lead to a surge in erroneously rejected applications and renewals, putting coverage
at risk for people who are eligible for Medicaid (Rosenbaum, Handley, and Morris 2021). The Centers
for Medicare & Medicaid Services recently issued updated guidance stating Medicaid eligibility and
enrollment backlogs should be processed within 12 months of the end of the public health
emergency.18 The guidance also prohibits states from terminating Medicaid coverage for people
deemed ineligible during the public health emergency until the state has completed an additional
redetermination of eligibility after the emergency ends. Finally, under previous guidance from
December 2020, the Centers for Medicare & Medicaid Services expected states to prioritize eligibility
and enrollment actions for people most likely to no longer be eligible for coverage (Musumeci and
Dolan 2021). The updated guidance requires states to consider how their approaches for processing
these actions will ensure continuity of coverage for eligible people and limit delays for those who
become newly eligible. State officials can begin preparing for the end of the public health emergency
now and avoid terminating coverage based on outdated information for eligible enrollees, many of
whom experienced disruptions to their employment and housing during the pandemic (Wagner 2020).
Medicaid enrollees whose incomes have risen above the eligibility threshold in their state will no
longer qualify for coverage when the disenrollment freeze expires. If such adults lack access to
affordable ESI, they will need to turn to the private nongroup market to remain insured. The
temporarily expanded Marketplace premium tax credits under the American Rescue Plan Act will make
Marketplace plans more affordable, but some adults may not be aware of the availability of zero-
premium or low-cost plans. Outreach and enrollment assistance can help adults transition from
Medicaid to Marketplace coverage and avoid disruptions in care (Haley and Wengle 2021). State
agencies will also need to assess eligibility for subsidized Marketplace coverage and other insurance
8 UNINSURANCE HELD STEADY , PUBLIC COVERAGE INCREASED DURING THE PANDEMICaffordability programs for adults who lose Medicaid eligibility after the public health emergency ends
(Musumeci and Dolan 2021).
The American Rescue Plan Act increased the subsidy amounts of Marketplace premium tax
credits, reducing the percentage of income people have to pay toward premiums, and expanded
eligibility for premium tax credits to adults with incomes above 400 percent of FPL. If Congress does
not extend these changes, they will expire at the end of 2022. Making the enhanced subsidies
permanent could reduce the number of people uninsured in the longer term, and most of the coverage
gains would occur among adults with moderate incomes (Banthin et al. 2021).
Policymakers can further reduce uninsurance by addressing the high uninsurance rates among
adults with low incomes, particularly in the remaining Medicaid nonexpansion states, where more than
one-third of adults with incomes at or below 138 percent of FPL are uninsured. The American Rescue
Plan Act provides these states with new incentives to expand Medicaid by increasing the federal
matching rate for regular (i.e., nonexpansion) Medicaid populations for two years (Musumeci 2021). If
the nonexpansion states had adopted Medicaid expansion in 2020, 4.4 million fewer people would
have been uninsured that year (Buettgens 2021). Federal policymakers are also considering
approaches for closing the Medicaid coverage gap in states that have not expanded eligibility under
the ACA.19
Additional health care reforms, ranging from incremental improvements to the ACA to more
comprehensive approaches, can advance the US toward universal coverage, though they have
different trade-offs in costs, provider payment rates, and disruptions to the existing health care system
(Blumberg et al. 2019).
Data and Methods
This brief draws on data from the Urban Institute’s Health Reform Monitoring Survey, a nationally
representative, internet-based survey of adults ages 18 to 64. Launched in 2013, the HRMS provides
timely information on health insurance coverage, health care access and affordability, and other health
topics before federal survey data become available. For each round of the HRMS, we draw a stratified,
random sample of nonelderly adults from Ipsos’s KnowledgePanel, the nation’s largest probability-
based online panel. Members of the panel are recruited from an address-based sampling frame
covering approximately 97 percent of US households, including those without internet access. If
needed, panel members are given internet access and web-enabled devices to facilitate their
participation.
For this analysis, we used data from the March 2019, March/April 2020, and April 2021 rounds of
the HRMS. The 2019 round was fielded March 4 through 14; it had a sample size of 9,596 adults, and
91 percent completed the survey in the first week of fielding. The 2020 round was fielded March 25
through April 10; it had a sample size of 9,032 adults, and 75 percent completed the survey in the first
week. And the 2021 round was fielded April 2 through 20; it had a sample size of 9,067 adults, and 82
percent completed the survey in the first week.
UNINSURANCE HELD STEADY, PUBLIC COVERAGE INCREASED DURING THE PANDEMIC 9The 2019 round of the HRMS included an oversample of adults with incomes below 138 percent
of FPL. In 2020, we changed the survey’s design to include larger oversamples of adults in low- and
moderate-income households, nonwhite and Hispanic/Latinx adults, and young adults. Survey weights
adjust for unequal selection probabilities and are poststratified to the characteristics of the national
nonelderly adult population, based on benchmarks from the Current Population Survey and the
American Community Survey. Participants can take the survey in English or Spanish, and the survey
takes a median of 15 minutes to complete. The margin of sampling error, including the design effect,
for the full sample of adults in the 2021 survey round is plus or minus 1.2 percentage points for a 50
percent statistic at the 95 percent confidence level.
Health Insurance Coverage Measures
In all rounds of the HRMS, respondents received a question, adapted from the American Community
Survey, about their current health insurance coverage. Respondents could report more than one type
of coverage, and those who did not report any coverage were asked to verify if they have health
insurance. We used additional follow-up questions to determine whether respondents enrolled in their
health plan through the Marketplace, whether they enrolled in a private plan through the Marketplace,
whether they are covered under certain state programs, and the name of the health plan for their main
source of coverage.
Because respondents could report more than one coverage type, we established a hierarchy of
responses to assign coverage types so that coverage estimates sum to 100 percent: ESI/military
coverage; public coverage, including Medicare, Medicaid, and CHIP; private nongroup coverage
purchased through or outside the Marketplaces; and other unspecified coverage. To address the
challenges associated with identifying health insurance coverage type in surveys (Call et al. 2013;
Klerman et al. 2009; Pascale 2008; Pascale, Fertig, and Call 2019), we used a logical editing process to
identify the most likely type of health insurance coverage held by respondents, based on the
information they provided in the survey (Blavin, Karpman, and Zuckerman 2016). However,
measurement error still occurs in survey estimates of coverage type, particularly in reports of private
nongroup coverage (which can be purchased through government-run Marketplaces with public
subsidies) and Medicaid coverage (which is often provided through private Medicaid managed-care
plans).
Estimates from this brief are not directly comparable with estimates from HRMS analyses from
before 2020 because of a change in the coverage editing process for respondents who reported
having insurance but did not report a specific coverage type and who did not enroll in a health plan
through the Marketplace. Under the previous approach, these respondents were identified as insured
with an unspecified coverage type if they reported having a deductible. The updated approach only
assigns unspecified coverage to these respondents if they report the name of a health plan that
provides a valid form of comprehensive health insurance coverage. Based on this update, respondents
reporting plans that do not offer comprehensive health insurance (e.g., health care sharing ministries)
are considered uninsured, yielding slightly higher estimates of uninsurance in this brief than in
10 UNINSURANCE HELD STEADY , PUBLIC COVERAGE INCREASED DURING THE PANDEMICprevious analyses of the HRMS. Under this updated coverage editing approach, estimates of the share of uninsured nonelderly adults in previous rounds of the HRMS would be 1 to 2 percentage points higher than under the previous approach. We applied the updated coverage editing process consistently for all years of data in this brief. Analysis Estimated changes in coverage are regression adjusted to control for any changes in the demographic and socioeconomic characteristics of respondents in each survey round not fully captured in the survey weights. This allows us to remove variation in coverage caused by changes in the observable characteristics of people responding to the survey over time. We control for measures used in poststratification of both the KnowledgePanel and the HRMS, including gender, age, race and ethnicity, primary language, educational attainment, marital status, presence of children in the household, household income, family income, homeownership status, internet access, urban/rural residence, and region. We also control for citizenship status and participation in the previous round of the survey. In presenting the regression-adjusted estimates, we use the predicted rate of each coverage measure in each year for the same nationally representative population. For this analysis, we base the nationally representative sample on respondents for the 2020 and 2021 rounds of the survey. We emphasize changes in coverage that are statistically different from 0 at the 5 percent level or lower and provide a 95 percent confidence interval for key estimates of changes in the number of adults with selected coverage types. Limitations This analysis has several limitations. First, studies have found significant measurement error in reported health insurance coverage type across surveys (Call et al. 2013; Klerman et al. 2009; Pascale 2008; Pascale, Fertig, and Call 2019). We attempt to mitigate this error using a logical editing process for coverage type that relies on multiple data elements (Blavin, Karpman, and Zuckerman 2016). Second, the probability-based internet panel underlying the HRMS does not cover some adult populations, including those who are homeless, are institutionalized, or do not speak English or Spanish. Third, the HRMS has a low cumulative response rate, and nonresponse bias is likely only partially mitigated by the survey weights. However, previous studies assessing recruitment for the panel from which HRMS samples are drawn have found little evidence of nonresponse bias for core demographic and socioeconomic measures (Garrett, Dennis, and DiSogra 2010; Heeren et al. 2008). Further, HRMS estimates of changes in coverage have been consistent with estimates from federal surveys with larger samples sizes, higher response rates, and stronger designs (Karpman and Long 2015). Finally, though nonresponse in federal surveys increased significantly during the pandemic (Dahlhamer et al. 2021; Rothbaum and Bee 2021), we find little change in nonresponse in the HRMS. Probability-based internet panels could potentially have more stable response patterns because panel members have previously agreed to participate in surveys. However, the impact of the pandemic on these types of surveys is not yet fully understood. UNINSURANCE HELD STEADY, PUBLIC COVERAGE INCREASED DURING THE PANDEMIC 11
Notes
1 The Families First Coronavirus Response Act has provided all states with a temporary increase in federal
matching funds for Medicaid beneficiaries not in the ACA Medicaid expansion population. To receive the
higher rate, states must follow several maintenance-of-effort requirements, including not disenrolling people
from Medicaid unless they request termination of coverage or move to a different state. These provisions will
remain in place at least until the end of the calendar quarter when the secretary of health and human services
declares the end of the public health emergency.
2 The states that did not expand Medicaid by April 2021 are Alabama, Georgia, Florida, Kansas, Mississippi,
Missouri, North Carolina, Oklahoma, South Carolina, South Dakota, Tennessee, Texas, Wisconsin, and
Wyoming. Wisconsin has used state funding to expand eligibility to nonelderly adults with incomes up to the
FPL. In other nonexpansion states, parents generally must have very low incomes to qualify for Medicaid, and
nonpregnant, nondisabled adults who are not parents living with dependent children are ineligible. In 2020,
voters in Missouri and Oklahoma approved ballot initiatives to expand Medicaid by July 1, 2021. Oklahoma’s
expansion took effect as scheduled. However, the Missouri legislature did not provide funding for the
expansion in the state budget, and the governor withdrew the state plan amendment for the expansion. On
July 22, 2021, the Missouri Supreme Court ruled that the state must implement the Medicaid expansion. For
this analysis, we treat Missouri and Oklahoma as nonexpansion states because they did not implement their
expansions by April 2021.
3 Under the American Rescue Plan Act, many people with incomes above 400 percent of FPL are eligible for
premium tax credits, but expanded eligibility is set to expire after 2022.
4 Coverage estimates often vary across surveys because of differences in survey design (Au-Yeung and Hest
2019). In this brief, we discuss statistically significant changes in coverage over the study period. Previous
analyses have found HRMS estimates of coverage changes to be consistent with estimates from other surveys
(Karpman and Long 2015).
5 US Bureau of Labor Statistics, “The Employment Situation – May 2021,” news release, June 4, 2021,
https://www.bls.gov/news.release/pdf/empsit.pdf.
6 We multiplied the estimated 2.7 percentage-point change in ESI between March 2019 and April 2021 by the
projected number of adults ages 18 to 64 in 2021. We used national population predictions from the US
Census Bureau stratified by race, ethnicity, and sex for people of all ages from 2016 to 2060, based on
estimated birth, death, and net migration rates over the period. Using the “main series” file, we summed the
2021 population projections for all nonelderly adults to arrive at 203,018,143 such adults that year. See “2017
National Population Projections Datasets,” US Census Bureau, February 20, 2020,
https://www.census.gov/data/datasets/2017/demo/popproj/2017-popproj.html.
7 In this brief, we combine Medicare, Medicaid, CHIP, and other government- or state-sponsored health plans
into a single measure of public coverage because survey respondents may confuse the names of these
coverage types (Pascale 2008). For a previous fact sheet based on data from the March/April 2020 HRMS and
the Urban Institute’s September 2020 Coronavirus Tracking Survey, we excluded Medicare from estimated
changes in public coverage (Karpman and Zuckerman 2020). Estimates in this brief also differ slightly from
estimates in that analysis because of differences in the survey weights and the regression adjustment, which
we describe in the Data and Methods section.
8
Administrative data show an increase of approximately 6 million adults enrolled in Medicaid between February
2020 and January 2021 in the 49 states and DC that report adult and child enrollment separately (Corallo and
Rudowitz 2021). Differences between the HRMS estimates of changes in public coverage and administrative
data for Medicaid enrollment may reflect several factors, including differences in the study period; inclusion of
18-year-olds as adults in the HRMS; inclusion of Medicare, CHIP, and state programs other than Medicaid in
the definition of public coverage in the HRMS; survey sampling error; and measurement error in coverage type
reported in the survey.
9 The number of people selecting Marketplace plans increased from 11.4 million during the 2019 open
enrollment period (November 1–December 15, 2018) to approximately 12 million during the 2021 open
enrollment period (November 1–December 15, 2020). The Centers for Medicare & Medicaid Services reported
12 UNINSURANCE HELD STEADY , PUBLIC COVERAGE INCREASED DURING THE PANDEMICan additional 940,000 people enrolled in Marketplace coverage during the special enrollment period between
February 15 and April 30, 2021, compared with 266,000 and 391,000 people who signed up through special
enrollment periods based on qualifying life events during the same periods in 2019 and 2020. Though the 2021
special enrollment period was extended to August 15, about half of new enrollment during the period’s original
time frame (February 15–April 30, 2021) occurred in April. Thus, some of these enrollments may have occurred
after the HRMS was fielded. See “2021 Open Enrollment Report,” Centers for Medicare & Medicaid Services,
accessed June 30, 2021, https://www.cms.gov/files/document/health-insurance-exchanges-2021-open-
enrollment-report-final.pdf; and “2021 Marketplace Special Enrollment Report,” Centers for Medicare &
Medicaid Services, May 6, 2021, https://www.cms.gov/newsroom/fact-sheets/2021-marketplace-special-
enrollment-period-report-1.
10 The shares of adults with an unspecified coverage type were 2.3 percent in 2019, 1.4 percent in 2020, and 1.3
percent in 2021.
11 Joan Alker and Allie Corcoran, “What Is Happening with Medicaid Enrollment in Q1 of 2021?” Say Ahhh! (blog),
Georgetown University Health Policy Institute, Center for Children and Families, May 21, 2021,
https://ccf.georgetown.edu/2021/05/21/what-is-happening-with-medicaid-enrollment-in-q1-of-2021/.
12 Noncitizens’ eligibility for Medicaid depends on several factors, including whether they are lawfully present,
considered qualified noncitizens based on their immigration status, and subject to the five-year waiting period
after receiving qualified status. See “Coverage for Lawfully Present Immigrants,” Centers for Medicare &
Medicaid Services, accessed June 30, 2021, https://www.healthcare.gov/immigrants/lawfully-present-
immigrants/.
13 “State Health Facts: Medicaid and CHIP,” Kaiser Family Foundation, accessed June 30, 2021,
https://www.kff.org/state-category/medicaid-chip/medicaidchip-eligibility-limits/.
14 The increase in public coverage between 2019 and 2021 in nonexpansion states was concentrated among the
group of adults most likely to be eligible for Medicaid or CHIP: 18-year-olds (who qualify for Medicaid or CHIP
based on eligibility thresholds for children), adults living with children under 18 in the household (who
potentially qualify as parents or caregivers), and adults in Wisconsin, which has used state funds to provide
coverage to adults with incomes up to the FPL (data not shown). The increase in public coverage for other
adults was statistically significant but small in magnitude.
15 “Opportunity Insights Economic Tracker,” Harvard University, accessed July 14, 2021,
https://www.tracktherecovery.org/.
16 Alker and Corcoran, “What Is Happening with Medicaid Enrollment in Q1 of 2021?” Say Ahh!.
17 Norris Cochran (acting secretary, US Department of Health and Human Services), letter to governors regarding
the public health emergency, January 22, 2021, https://ccf.georgetown.edu/wp-
content/uploads/2021/01/Public-Health-Emergency-Message-to-Governors.pdf.
18 Daniel Tsai (deputy administrator and director, Centers for Medicare & Medicaid Services), letter to state
health officials regarding, “Updated Guidance Related to Planning for the Resumption of Normal State
Medicaid, Children’s Health Insurance Program (CHIP), and Basic Health Program (BHP) Operations upon
Conclusion of the COVID-19 Public Health Emergency,” August 13, 2021, https://www.medicaid.gov/federal-
policy-guidance/downloads/sho-21-002.pdf.
19 Rachel Roubein and Alice Miranda Ollstein, “Plugging Obamacare’s Biggest Hole Poses Dilemma for
Democrats,” Politico, July 10, 2021, https://www.politico.com/news/2021/07/10/obamacare-medicaid-
coverage-gap-democrats-499013.
UNINSURANCE HELD STEADY, PUBLIC COVERAGE INCREASED DURING THE PANDEMIC 13References
ASPE (US Department of Health and Human Services, Office of the Assistant Secretary for Planning and
Evaluation). 2021. Health Coverage under the Affordable Care Act: Enrollment Trends and State Estimates.
Washington, DC: US Department of Health and Human Services, Office of the Assistant Secretary for Planning
and Evaluation.
Au-Yeung, Caroline, and Robert Hest. 2019. “Comparing Federal Government Surveys That Count the Uninsured:
2019.” Minneapolis: State Health Access Data Assistance Center.
Banthin, Jessica, Matthew Buettgens, Michael Simpson, and Robin Wang. 2021. “What If the American Rescue
Plan’s Enhanced Marketplace Subsidies Were Made Permanent? Estimates for 2022.” Washington, DC: Urban
Institute.
Banthin, Jessica, Michael Simpson, Matthew Buettgens, Robin Wang, and Linda J. Blumberg. 2020. “Changes in
Health Insurance Coverage Due to the COVID-19 Recession: Preliminary Estimates Using Microsimulation.”
Washington, DC: Urban Institute.
Blavin, Fredric, Michael Karpman, and Stephen Zuckerman. 2016. “Understanding Characteristics of Likely
Marketplace Enrollees and How They Choose Plans.” Health Affairs 35 (3): 535–39.
https://doi.org/10.1377/hlthaff.2015.0867.
Blumberg, Linda J., John Holahan, Matthew Buettgens, Anuj Gangopadhyaya, Bowen Garrett, Adele Shartzer,
Michael Simpson, Robin Wang, Melissa Favreault, and Diane Arnos. 2019. From Incremental to Comprehensive
Health Reform: How Various Reform Options Compare on Coverage and Costs. Washington, DC: Urban Institute.
Brooks, Tricia, and Andy Schneider. 2020. “The Families First Coronavirus Response Act: Medicaid and CHIP
Provisions Explained.” Washington, DC: Georgetown University Health Policy Institute, Center for Children and
Families.
Buettgens, Matthew. 2021. “Medicaid Expansion Would Have a Larger Impact Than Ever during the COVID-19
Pandemic.” Washington, DC: Urban Institute.
Call, Kathleen T., Michael E. Davern, Jacob A. Klerman, and Victoria Lynch. 2013. “Comparing Errors in Medicaid
Reporting across Surveys: Evidence to Date.” Health Services Research 48 (2 pt. 1): 652–64.
https://dx.doi.org/10.1111%2Fj.1475-6773.2012.01446.x.
Corallo, Bradley, and Robin Rudowitz. 2021. “Analysis of Recent National Trends in Medicaid and CHIP
Enrollment.” San Francisco: Kaiser Family Foundation.
Dahlhamer, James M., Matthew D. Bramlett, Aaron Maitland, and Stephen J. Blumberg. 2021. “Preliminary
Evaluation of Nonresponse Bias Due to the COVID-19 Pandemic on National Health Interview Survey
Estimates, April–June 2020.” Hyattsville, MD: Centers for Disease Control and Prevention, National Center for
Health Statistics.
Garfield, Rachel, Gary Claxton, Anthony Damico, and Larry Levitt. 2020. “Eligibility for ACA Health Coverage
Following Job Loss.” San Francisco: Kaiser Family Foundation.
Garrett, Bowen, and Anuj Gangopadhyaya. 2020. “How the COVID-19 Recession Could Affect Health Insurance
Coverage.” Washington, DC: Urban Institute.
Garrett, Joe, J. Michael Dennis, and Charles A. DiSogra. 2010. “Non-response Bias: Recent Findings from Address-
Based Panel Recruitment.” Presented at the Annual Conference of the American Association for Public Opinion
Research, Chicago, May 13–16.
Haley, Jennifer M., and Erik Wengle. 2021. “Many Uninsured Adults Have Not Tried to Enroll in Medicaid or
Marketplace Coverage.” Washington, DC: Urban Institute.
Heeren, Timothy, Erika M. Edwards, J. Michael Dennis, Sergei Rodkin, Ralph W. Hingson, and David L.
Rosenbloom. 2008. “A Comparison of Results from an Alcohol Survey of a Prerecruited Internet Panel and the
National Epidemiologic Survey on Alcohol and Related Conditions.” Alcoholism: Clinical and Experimental
Research 32 (2): 222–29. https://doi.org/10.1111/j.1530-0277.2007.00571.x.
14 UNINSURANCE HELD STEADY , PUBLIC COVERAGE INCREASED DURING THE PANDEMICHolahan, John, and Vicki Chen. 2011. “Changes in Health Insurance Coverage in the Great Recession.” Washington, DC: Kaiser Commission on Medicaid and the Uninsured. Karpman, Michael, and Sharon K. Long. 2015. “QuickTake: HRMS Benchmarks Well Against Gallup-Healthways and NHIS on Changes in Health Insurance Coverage between 2013 and 2015.” Washington, DC: Urban Institute. Karpman, Michael, and Stephen Zuckerman. 2020. “ACA Offers Protection as the COVID-19 Pandemic Erodes Employer Health Insurance Coverage.” Washington, DC: Urban Institute. Karpman, Michael, Stephen Zuckerman, and Genevieve M. Kenney. 2020. “Uneven Recovery Leaves Many Hispanic, Black, and Low-Income Adults Struggling.” Washington, DC: Urban Institute. Khorrami, Peggah, and Benjamin D. Sommers. 2021. “Changes in US Medicaid Enrollment during the COVID-19 Pandemic.” JAMA Network Open 4 (5): e219463. https://doi.org/10.1001/jamanetworkopen.2021.9463. Klerman, Jacob A., Michael Davern, Kathleen T. Call, Victoria Lynch, and Jeanne D. Ringel. 2009. “Understanding the Current Population Survey’s Insurance Estimates and the Medicaid ‘Undercount.’” Health Affairs 28 (6): w991–w1001. https://doi.org/10.1377/hlthaff.28.6.w991. Long, Sharon K., Genevieve M. Kenney, Stephen Zuckerman, Dana E. Goin, Douglas Wissoker, Fredric Blavin, Linda J. Blumberg, Lisa Clemans-Cope, John Holahan, and Katherine Hempstead. 2014. “The Health Reform Monitoring Survey: Addressing Data Gaps to Provide Timely Insights into the Affordable Care Act.” Health Affairs 33 (1): 161–67. https://doi.org/10.1377/hlthaff.2013.0934. Musumeci, MaryBeth. 2021. “Medicaid Provisions in the American Rescue Plan Act.” San Francisco: Kaiser Family Foundation. Musumeci, MaryBeth, and Rachel Dolan. 2021. “Key Issues for State Medicaid Programs When the COVID-19 Public Health Emergency Ends.” San Francisco: Kaiser Family Foundation. Obama, Barack. 2016. “United States Health Reform: Progress and Next Steps.” JAMA 316 (5): 525–32. https://doi.org/10.1001/jama.2016.9797. Pascale, Joanne. 2008. “Measurement Error in Health Insurance Reporting.” Inquiry 45 (4): 422–37. https://doi.org/10.5034%2Finquiryjrnl_45.04.422. Pascale, Joanne, Angela R. Fertig, and Kathleen T. Call. 2019. “Assessing the Accuracy of Survey Reports of Health Insurance Coverage Using Enrollment Data.” Health Services Research 54 (5): 1099–109. https://doi.org/10.1111/1475-6773.13191. Rosenbaum, Sara, Morgan Handley, and Rebecca Morris. 2021. “Winding Down Continuous Enrollment for Medicaid Beneficiaries When the Public Health Emergency Ends.” New York: Commonwealth Fund. Rothbaum, Jonathan, and Adam Bee. 2021. “Coronavirus Infects Surveys, Too: Survey Nonresponse Bias and the Coronavirus Pandemic.” Washington, DC: US Census Bureau. Wagner, Jennifer. 2020. “States Can Act Now to Keep Medicaid Enrollees Covered When the Public Health Emergency Ends.” Washington, DC: Center on Budget and Policy Priorities. UNINSURANCE HELD STEADY, PUBLIC COVERAGE INCREASED DURING THE PANDEMIC 15
About the Authors
Michael Karpman is a senior research associate in the Health Policy Center at the Urban Institute. His
work focuses primarily on the implications of the Affordable Care Act, including quantitative analysis
related to health insurance coverage, access to and affordability of health care, use of health care
services, and health status. His work includes overseeing and analyzing data from the Urban Institute’s
Health Reform Monitoring Survey and Well-Being and Basic Needs Survey. Before joining Urban in
2013, Karpman was a senior associate at the National League of Cities Institute for Youth, Education,
and Families. He received his MPP from Georgetown University.
Stephen Zuckerman is a senior fellow and vice president for health policy at the Urban Institute. He
has studied health economics and health policy for 30 years and is a national expert on Medicare and
Medicaid physician payment, including how payments affect enrollee access to care and the volume of
services they receive. He is currently examining how payment and delivery system reforms can affect
the availability of primary care services and studying the implementation and impact of the Affordable
Care Act. Before joining Urban, Zuckerman worked at the American Medical Association’s Center for
Health Policy Research. He received his PhD in economics from Columbia University.
16 UNINSURANCE HELD STEADY , PUBLIC COVERAGE INCREASED DURING THE PANDEMICAcknowledgments
This brief was funded by the Robert Wood Johnson Foundation. The views expressed here do not
necessarily reflect the views of the Foundation.
The views expressed are those of the authors and should not be attributed to the Urban Institute,
its trustees, or its funders. Funders do not determine research findings or the insights and
recommendations of Urban experts. Further information on the Urban Institute’s funding principles is
available at urban.org/fundingprinciples.
The authors gratefully acknowledge Linda J. Blumberg, Matthew Buettgens, and Jennifer M. Haley
for helpful feedback and Rachel Kenney for her careful editing.
ABOUT THE URBAN INSTITUTE
The nonprofit Urban Institute is a leading research organization dedicated to
developing evidence-based insights that improve people’s lives and strengthen
communities. For 50 years, Urban has been the trusted source for rigorous analysis
of complex social and economic issues; strategic advice to policymakers,
philanthropists, and practitioners; and new, promising ideas that expand
opportunities for all. Our work inspires effective decisions that advance fairness
500 L’Enfant Plaza SW
and enhance the well-being of people and places.
Washington, DC 20024
Copyright © August 2021. Urban Institute. Permission is granted for reproduction
www.urban.org
of this file, with attribution to the Urban Institute.
UNINSURANCE HELD STEADY, PUBLIC COVERAGE INCREASED DURING THE PANDEMIC 17You can also read
